Abstract We are in an exciting era of biology where the inner workings of cells can be explored by rapidly developing imaging methods. Fluorescence microscopy has two major advantages: labeling specificity and live cell compatibility. However, it is limited by diffraction to approximately 250 nm resolution. The recent advent of single molecule switching nanoscopy (SMSN, also known as PALM/STORM/FPALM) has overcome this fundamental limit by stochastically switching single dyes on and off such that their emission events are separated in time. This allows their center positions to be localized with high precision in space leading to a reconstructed super resolved image with a resolution down to ~25 nm. However, its biological application is limited for two reasons: (1) SMSN applications are typically limited to fixed samples due to the poor temporal resolution and (2) the application been limited to structures close to the coverslip in thin samples because of its inferior resolution in the depth direction (z) and rapidly deteriorating resolution in thick samples. Further, SMSN generates thousands to millions of precise single molecule positions per dataset - a large amount of information rarely explored due to the lack of data quantification methods. Overcoming these hurdles will allow visualization and quantification of nanostructures in living cells, determine the stoichiometry of fluorescently tagged proteins and thus drastically expand the breadth of SMSN applications. We propose to (1) develop interferometric SMSN for ultra-high resolution imaging in live cells and thick samples capturing 3D live cell dynamics through an imaging depth up to 50 m with isotropic 5-10 nm resolution; (2) develop structure and stoichiometry mapping in space, time and multiple color to build high- resolution 3D models of macromolecular complexes and large protein assemblies in live cell; and (3) further improve the resolution by another order of magnitude (~1 nm precision) of the reconstructed model by a high- content system allowing statistical quantification over thousands of cells (~3000 cells per hour). Applying these developments, we will study the distinct molecular organization and function of three different myosins during cytokinesis in live fission yeast and neuronal motility focusing on the growth cones in live neuron. The proposed research will, for the first time, make ultra-high resolution visualization of cells possible in thick and live samples, allow building highly-resolved and evolving structure and stoichiometry models of macromolecular assemblies and protein clusters in vivo and further categorizing them based on their live-cell context. This allows us to determine the organization of myosin molecules in vivo, visualize their interaction with actin network and study their function in tension generation within the cytokinetic ring during cell division. The proposed research is enthusiastically supported by my close collaboration with Martin Booth, adaptive optics expert from Oxford University, Daniel Suter neuron biologist from Purdue University and Thomas Pollard, whose research focuses on molecular basis of cellular motility and cytokinesis from Yale University.